Nuthawut - stock.adobe.com
The concept of virtualizing edge devices isn't new. VMware demonstrated ESXi running on an Advanced RISC Machine architecture back in 2018. However, it hasn't been released yet or even put out for testing. VMware intends to gauge interest in the technology and look for hardware partners to support it and develop meaningful use cases before making it generally available.
Traditionally, a centralized data center contained enterprise computing and storage resources. The data center housed the servers, and client systems outside the data center accessed the storage and server-based applications. This model worked for decades, but the computing needs of modern organizations have led to a fundamental change in architecture for large-scale IT infrastructures.
Why move to the edge?
With more applications, more users, greater operational scope and unimaginably vast quantities of data -- much of which must be generated and handled in real-time -- networks now handle severe, massive loads. Network bandwidth limitations, latencies and congestion or disruption can all have terrible consequences for organizations.
Some organizations seek solutions to these problems at the edge of the infrastructure rather than in the data center itself. By moving computing out of principal data centers and closer to the locations of data collection, organizations can create small local networks at remote installations and eliminate many of the limitations and uncertainties of today's internet.
This means the organization must build compute facilities at each remote location. Such edge data center are typically quite small compared to traditional facilities and require very little power and cooling. They often take up no more space than an average closet.
Edge data center types
An edge data center generally falls into one of two categories: micro data centers and nano data centers.
A micro data center is essentially a scaled-down version of a traditional data center. The micro facility usually hosts one or more rack-mounted servers capable of running thousands of VMs or containers. The facility typically has high integration -- such as a hyper-converged infrastructure -- and high availability. It can operate with a high degree of software-defined capability and uses software-defined storage and software-defined networking. Ultimately, the micro data center can handle a rich set of enterprise-level services and applications. Organizations can deploy micro data centers in shipping containers or workspace closets.
A nano data center is usually much smaller and far more limited than a micro data center. Nano data centers might use ruggedized equipment -- such as computers, storage and devices designed for operation in non-IT conditions and capable of withstanding environmental factors such as vibration, heat and cold -- capable of supporting a small subset of services and applications across less than 100 VMs or containers. An organization might deploy a nano data center made up of little more than one or two servers in a protected enclosure in the base of a cell tower or boxed-in a rail yard. Most nano data centers exist in purpose-specific environments.
Advanced RISC Machine processors don't define micro and nano data centers, but many micro and nano data center environments employ devices with ARM or other RISC processors and can eventually run ESXi on those processors. IoT can employ up to hundreds of sensors to collect data at an industrial location such as a power plant or manufacturing facility. The sensors must use ARM -- or other RISC -- processors to pass data across a local network. That data then goes to a small number of servers that locally store, process and normalize the data before passing the processed data to the main data center. This is the most common method for nano data center deployment.
Dig Deeper on Remote administrative options
Related Q&A from Stephen J. Bigelow
Containers have rapidly come into focus as a popular option for deploying applications, but they have limitations and are fundamentally different ... Continue Reading
ALM and SDLC both cover much of the same ground, such as development, testing and deployment. Where these lifecycle concepts differ is the scope of ... Continue Reading
Eliciting performance requirements from business end users necessitates a clearly defined scope and the right set of questions. Expert Mary Gorman ... Continue Reading